SUMMARY Alzheimer?s disease and related dementias (ADRDs) are typically the result of neurodegenerative processes that begin 15-20 years before clinical diagnosis. Despite this lengthy subclinical phase and advances in early diagnostic technologies including cerebrospinal fluid (CSF) biomarkers, brain MRIs, and positron emission tomography (PET) measures there are no scalable biomarkers to identify individuals with preclinical disease, which may offer the best time window for intervention. Specifically, because brain MRI and PET tests are resource-intensive and CSF sampling is invasive, there is urgent need for biomarkers based on blood samples, which can be non-invasively collected even in centers lacking highly specialized technology. The mitochondrial genome (mtDNA), unlike the stable nuclear genome, is dynamic and accumulates somatic mutations over the lifespan. Mutations in the mtDNA can be induced by oxidative stress and lead to declining mitochondrial function as we age; in turn, less functional mitochondria generate higher levels of oxidative stress, which induces chronic inflammation and tissue injury in specific organs including the brain. Yet, no study has investigated mtDNA mutations as early predictors of ADRD. We hypothesize that individuals with higher levels and faster accumulation of blood mtDNA mutations have greater cognitive decline and preclinical ADRD brain imaging changes during midlife, and that these mtDNA biomarkers will predict ADRD diagnosis in older adults. We will test our hypotheses by leveraging the NHLBI-funded Coronary Artery Risk Development In young Adults (CARDIA) study. CARDIA is a multicenter, community-based, longitudinal cohort that recruited 5,115 black and white young adults (mean age 25 years), followed them up at least every five years, and is preparing to conduct its examination Year 35 (Y35) visit in 2020/21 (mean age 60 years) when over 3,000 participants are expected to return. We will use state-of-the-art deep sequencing technology to measure mtDNA mutations in CARDIA blood samples collected at Y15, Y25, and Y35. In Aim 1, we will determine whether higher levels of mtDNA mutations and their accumulation over time are associated with greater cognitive decline in midlife. In Aim 2, we will determine whether mtDNA mutations are associated with structural, physiological, and functional MRI phenotypes of preclinical ADRD as well as with sensitive MRI-based markers of accelerated brain aging and preclinical ADRD constructed by our team using contemporary machine learning techniques. Finally, in Aim 3, we will test the clinical utility of these mtDNA mutation biomarkers in identifying individuals at risk of future ADRD in four older cohorts that have large numbers of longitudinally identified, clinically-diagnosed ADRD. Our focus on longitudinal measures of blood mtDNA, cognitive function, and MRI changes over a 10- year period during midlife, combined with characterization of their clinical utility in older populations, is highly innovative. If successful, our study may help initiate possible future interventions in early midlife and reduce the public health and medical burden of AD and dementia.